Presented at the Neonatal Society 2006 Widdowson Meeting.
Burns C1, Counsell S2, Allsop J2, Cowan F1, Rutherford M2
1 Department of Paediatrics, Imperial College, Hammersmith Hospital, London, UK
2 Imaging Sciences, Imperial College, Hammersmith Hospital, London, UK
Background: Neonatal hypoglycaemia is a major clinical problem and a cause of long-term neurological morbidity and occasional mortality. The most commonly reported pattern of brain injury localises to the occipital and parietal white matter (WM) and cortex (1). There is no good explanation for this posterior predilection and this data is mainly based on few magnetic resonance imaging (MRI) studies.
Aims: 1) to test the hypothesis that in term-born hypoglycaemic infants, the patterns of brain injury seen on MRI are more diverse than currently described 2) to determine any relationship between brain injury and the type (2), documented severity and duration of hypoglycaemia, 3) to investigate any antenatal or perinatal differences between hypoglycaemic neonates and neurologically normal low-risk control infants.
Methods: 35 patients were recruited. All were >36 weeks GA at birth, had at least one episode of hypoglycaemia (blood glucose ≤2.6mmol/l) and had an early (<6 weeks’) MRI brain scan. Infants with evidence of HIE were excluded. Detailed antenatal, perinatal and postnatal data were obtained and compared to equivalent data from a previously recruited cohort of 229 term-born control infants considered neurologically normal at birth and at 12-18 months (3). The patterns of brain lesions as seen on MR scans were correlated with the type, duration and severity of hypoglycaemia. Ethical permission for the study was obtained from the Hammersmith Hospital REC and permission for scanning obtained from the parents.
Results: 33 infants had abnormal scans. 94% had WM abnormality, which was classified as severe in 43%. The cortex was also involved in 49% with abnormal WM. Only 49% had a predominantly posterior pattern of injury. 31% had haemorrhagic lesions within the WM. 40% had basal ganglia (BG) lesions; these were often mild and only associated with an abnormal internal capsule in 11%. No particular pattern of injury was significantly related to the type, severity or duration of hypoglycaemia; however, multiple haemorrhagic lesions within the WM tended to be more strongly associated with severe or prolonged hypoglycaemia.
Pregnancy-induced hypertension (p=0.04) and family history of seizures (p=0.0008) were more common in cases than controls. Cases were less often born by normal vaginal delivery (p=0.01) and more often born by emergency caesarean (p≤0.0001). Case infants were more likely to require resuscitation at bir th (p=0.002). 14 infants were growth restricted and 1 macrosomic; 13 had prolonged hypoglycaemia resistant to treatment. 51% of case infants had known risk factors for hypoglycaemia. Only two mothers were diabetic.
Conclusions: We found the patterns of injury seen with neonatal hypoglycaemia more varied than described in the literature. WM injury was not confined to the posterior regions; haemorrhage has not been reported before and BG injury only rarely. Nearly half our cohort had no identifiable risk factors for hypoglycaemia. This study provides evidence for reviewing the current guidelines for screening of neonatal hypoglycaemia.
References
1. Alkalay AL et al. Clin Pediatr 2005; 44: 783-790.
2. Volpe JJ. Neurology of the newborn. 3 ed 1995.
3. Haataja L et al J Pediatrics 1999; 135: 153-61.